Wednesday, 6 November 2013

Marcel's Maxims

Marcel Böhme is a 5th year PhD student who now spends his days in sultry Singapore ‘breaking’ programs, as part of his research. Once upon a time, however, he grew up in a house at the edge of a forest in Schwepnitz, “a sleepy little village of 1500 souls in Germany.” He had an awesome, happy childhood growing up in the countryside, spending his winters walking through snow filled woods, following rabbit and deer prints, and his summers swimming in the nearby lake, chasing cows and falling from birch trees – which reads like lines straight off the pages of an Enid Blyton tale and something that, unfortunately, very few of us city-folk can relate to.

Photo credit: Nimantha Baranasuriya

What do you do now and where will you go from here?
My work is in automated software testing and debugging. Currently, I am looking at the changed behaviour of evolving programs. Let’s take Linux for example. The operating system has been evolving over the last twenty years to a massive 300 million lines of code and each day a ginormous 16 thousand lines of code are changed in the Kernel alone. I am interested whether any of these changes break anything, e.g., how you get a "blue screen". As part of my research, I have developed automated techniques that test and check these changes for errors. One of my tools found a major security problem in the Linux Core Utilities amongst other errors. Of course, there are many more interesting research questions that I'm looking into (e.g., automated debugging and program repair).

Hard to tell what my future plans are. I may follow the academic path for the time being. That gives intellectual freedom and opportunities to make some real impact. In the short term, it would be great to work with Professor Andreas Zeller in Germany. He has made significant contributions in my field of study and in particular to the automated debugging of large (evolving) software systems. In the long term, however, I still feel (geographically) unconstrained. Come what may, when I encounter an awesome opportunity, I'll go for it.

Describe your experience as an SoC student.
There are numerous opportunities to get involved and contribute. Earlier, I spent a year as a Graduate Student Representative and discussed with the school administration important topics related to the PhD program of SoC. Last year, I was invited to discussions with Singapore's Ministry of Education about possible changes to nationwide policies that were directly affecting us as PhD students. At some point, together with friends, we organized weekly PhD seminars (CSTalks) that fostered cross-lab interaction and gave direct feedback on interesting work in progress -- no advisors allowed. Next year, I am fortunate to meet Nobel Prize Laureates and Fields Medal Winners at the Global Young Researchers Summit (GYSS'14). Our PhD students are energetic and motivated. This liberal and competitive spirit drives high-quality research that is published in premier venues.

There is a lot of freedom in the way you can pursue your research. In Europe, I think PhD students often depend on grants from the industry and deliverables that totally need to work by the end of the year. Thus, there may be less opportunity for fundamental research which seeks longer term impact. In Singapore (specifically SoC), the majority of PhD students are funded on the Research Scholarship. So, without the pressure of immediate deliverables, you can do fundamental (farsighted, high payoff) research that won't expire in a few years with new technologies. But also, there are a lot of industrial grants and projects that can offer real data and relevant, applied research problems. For instance, the processor technology, developed by a friend during his PhD at SoC, has recently gained a lot of interest from ARM. Guess, where he is going to work very soon?

I also like the forward-mentality in SoC: If you just work hard enough, you can achieve something great. In a way, NUS (including SoC) is emitting this awesome sense of innovation, purpose, and progress. It is hard not to be inspired. If you have not already, you should go visit UTown and the CREATE laboratories to get a flavour.

What do you enjoy doing when you’re not working?
In my spare time, I am a jack of all trades, master of none. My philosophy is if other people can deeply appreciate doing something, then I can too. I like learning languages and spent a few semesters learning Italian and Mandarin Chinese. Unfortunately, not enough stuck with me to be able to converse. I like listening to classical music and attend mostly violin and piano concerti. In Dresden, I would pay a ridiculously small amount to visit the Semperoper as a student. In Singapore, the NUS Conservatory of Music organises classical concerti (free admission) and of course there is the Esplanade. I’m also the world’s worst guitar player and like taking courses on the online learning platform Courser.

I like to explore and choose the path least travelled. I've been to many places in South-East Asia and when I travel, I try to get lost to find my way back without a map. I also like to explore different cuisines and Singapore is *the* place to be for every food fanatic. Durian - damn shiok lah! Since my time is scarce, another means of exploration is by reading books. “A reader lives a thousand lives before he dies. The man who never reads lives only one.” -- George R. R. Martin. I cycle to school every day, read crime novels in the evening, and play badminton on weekends.

What is something most people would be surprised to learn about you?
My online poker winnings paid for my first flight to Singapore (though I never invested any real money). In the beginning, a friend gave me $10 to poker for fun. Upon closing the account, I cashed out more than a hundredfold. Chiefly, I learned how to manage risk and reputation in an aggressive and highly dynamic "market". However, I do strongly discourage any form of gambling!
[Ed: Ahem…]

What advice would you give a prospective PhD student? 
These are some maxims I picked up along the way:

Work hard but productively. Do more things in shorter time. Focus on progress. Working long hours alone, won't give you the edge.

Learn from your mentor. Draw from his or her experience. Prepare meetings. Listen first, then speak. Work out deliverables for the next meeting. Digest the feedback. Since everybody has to publish to an international community, the only factor differentiating an excellent from a good university is your research advisor.

Maximize Originality. If knowledge is a tree, don't work on the leaves. Work on thick branches that can continue to grow. You'll be known for that. Somebody else will take care of the leaves.

Make some noise. Invest in your academic reputation. Publish to premium venues. In conferences, don't hide. Introduce yourself and make your peers interested in your work. Visit other research teams.

Read a lot. Only then you can make new connections that nobody else made before. Follow the most important conferences, journals, and research celebrities in your domain.

Peel the Onion. Every onion has a concise problem statement at its core. Before starting research, reduce all that complexity that is not crucial to an intuitive solution of your problem statement. For instance, you work in program analysis -- don't get distracted by concurrency, object-orientation, and unsupported program constructs.

Elevator Pitch. Be able to explain problem, approach, and main results of your work (i.e., every paper and even your dissertation) in one minute! If you cannot explain your stuff quickly, concisely, and coherently, you won't convince your busy reviewers.

Do lots of sports. A healthy mind lives in a healthy body. I get the best ideas, when I away from the lab.

Best hawker stall in Singapore? 
JiaXiang, 721 Clementi West Ave 2. Delicious steamboat and Sichuan food.

Worst fear?
Is there a life after PhD?!

Three ultimate dinner party guests?
Nikola Tesla, Jean-Paul Sartre, and Kurt Gödel

Seriously, which faculty members' innermost thoughts and feelings are you most curious about? Email

No comments:

Post a Comment